Magnetic Resonance Angiography has great promise as a non-invasive method of obtaining information which now is only obtained with invasive angiography. However, of the several Magnetic Resonance Angiography (MRA) techniques now receiving serious attention from clinicians, almost all suffer from artifacts due to complex flow at the bifurcations and origins of arteries. It is well known that atherosclerotic plaques and aneurysms, which also occur at these locations, are associated with disturbed flow. Preliminary clinical experience suggests that artifacts associated with complex flow significantly reduce the reliability of MRA techniques. This proposal outlines a program for study and reduction of the artifacts associated with the complex flow in MRA. Computer calculations indicate that signals from regions with complicated flow interact with data acquisition techniques in specific ways to cause this type of signal loss. Hardware and software solutions are proposed. Additionally, a series of experiments on flow phantoms are proposed in which computer simulations are used to attempt to predict the source and the degree of signal loss. These simulations will deepen our understanding of the basic processes which give rise to this signal loss, which will continue to be the fundamental limit to the clinical utility of MRA for the foreseeable future.